SSRN-id3104847

Recommendations

Info

12 FINANCIAL MACHINE LEARNING AS A DISTINCT SUBJECT(c) 2018 by Marcos Lopez de Prado. Reprinted with permission. All rights reserved. Full version available at https://goo.gl/w6gMdq◦ Overfitting is unethical. It leads to promising outcomes that cannot be delivered.When done knowingly, overfitting is outright scientific fraud. The factthat many academics do it does not make it right: They are not risking anyone’swealth, not even theirs.◦ It is also a waste of your time, resources, and opportunities. Besides, theindustry only pays for out-of-sample returns. You will only succeed afteryou have created substantial wealth for your investors. How:◦ Chapters 11–15: There are three backtesting paradigms, of which historicalsimulation is only one. Each backtest is always overfit to some extent, and itis critical to learn to quantify by how much.◦ Chapter 16: Learn robust techniques for asset allocation that do not overfitin-sample signals at the expense of out-of-sample performance.1.3.3 Structure by Common PitfallDespite its many advantages, ML is no panacea. The flexibility and power of MLtechniques have a dark side. When misused, ML algorithms will confuse statisticalflukes with patterns. This fact, combined with the low signal-to-noise ratio thatcharacterizes finance, all but ensures that careless users will produce false discoveriesat an ever-greater speed. This book exposes some of the most pervasive errorsmade by ML experts when they apply their techniques on financial datasets. Some ofthese pitfalls are listed in Table 1.2, with solutions that are explained in the indicatedchapters.1.4 TARGET AUDIENCEThis book presents advanced ML methods specifically designed to address the challengesposed by financial datasets. By “advanced” I do not mean extremely difficult tograsp, or explaining the latest reincarnation of deep, recurrent, or convolutional neuralnetworks. Instead, the book answers questions that senior researchers, who haveexperience applying ML algorithms to financial problems, will recognize as critical.If you are new to ML, and you do not have experience working with complex algorithms,this book may not be for you (yet). Unless you have confronted in practice theproblems discussed in these chapters, you may have difficulty understanding the utilityof solving them. Before reading this book, you may want to study several excellentintroductory ML books published in recent years. I have listed a few of them in thereferences section.The core audience of this book is investment professionals with a strong ML background.My goals are that you monetize what you learn in this book, help us modernizefinance, and deliver actual value for investors.This book also targets data scientists who have successfully implemented MLalgorithms in a variety of fields outside finance. If you have worked at Google andhave applied deep neural networks to face recognition, but things do not seem toElectronic copy available at: https://ssrn.com/abstract=3104847
REQUISITES 13TABLE 1.2Common Pitfalls in Financial ML# Category Pitfall Solution Chapter(c) 2018 by Marcos Lopez de Prado. Reprinted with permission. All rights reserved. Full version available at https://goo.gl/w6gMdq1 Epistemological The Sisyphus paradigm The meta-strategyparadigm2 Epistemological Research through Feature importancebacktestinganalysis3 Data processing ChronologicalThe volume clock 2sampling4 Data processing Integer differentiation Fractionaldifferentiation55 Classification Fixed-time horizonlabeling6 Classification Learning side and sizesimultaneously7 Classification Weighting of non-IIDsamplesThe triple-barrier 3methodMeta-labeling 3Uniqueness weighting;sequentialbootstrapping8 Evaluation Cross-validationleakagePurging andembargoing9 Evaluation Walk-forwardCombinatorial purged(historical) backtesting cross-validation10 Evaluation Backtest overfitting Backtesting onsynthetic data; thedeflated Sharpe ratio1847,911,1210–16work so well when you run your algorithms on financial data, this book will helpyou. Sometimes you may not understand the financial rationale behind some structures(e.g., meta-labeling, the triple-barrier method, fracdiff), but bear with me: Onceyou have managed an investment portfolio long enough, the rules of the game willbecome clearer to you, along with the meaning of these chapters.1.5 REQUISITESInvestment management is one of the most multi-disciplinary areas of research, andthis book reflects that fact. Understanding the various sections requires a practicalknowledge of ML, market microstructure, portfolio management, mathematicalfinance, statistics, econometrics, linear algebra, convex optimization, discretemath, signal processing, information theory, object-oriented programming, parallelprocessing, and supercomputing.Python has become the de facto standard language for ML, and I have to assumethat you are an experienced developer. You must be familiar with scikit-learn(sklearn), pandas, numpy, scipy, multiprocessing, matplotlib and a few other libraries.Electronic copy available at: https://ssrn.com/abstract=3104847
Page 1 and 2: Electronic copy copy available at:
Page 3 and 4: Praise for Advances in Financial Ma
Page 5 and 6: machine learning. For academics and
Page 7 and 8: (c) 2018 by Marcos Lopez de Prado.
Page 17 and 18: CONTENTSxi(c) 2018 by Marcos Lopez
Page 19 and 20: CONTENTSxiiiExercises, 135Reference
Page 21 and 22: CONTENTSxvExercises, 208References,
Page 23 and 24: CONTENTSxvii(c) 2018 by Marcos Lope
Page 25 and 26: CONTENTSxix(c) 2018 by Marcos Lopez
Page 29 and 30: CHAPTER 1(c) 2018 by Marcos Lopez d
Page 31 and 32: THE MAIN REASON FINANCIAL MACHINE L
Page 33 and 34: BOOK STRUCTURE 7(c) 2018 by Marcos
Page 35 and 36: BOOK STRUCTURE 9(c) 2018 by Marcos
Page 37: BOOK STRUCTURE 11(c) 2018 by Marcos
Page 41 and 42: FAQs 15(c) 2018 by Marcos Lopez de
Page 43 and 44: FAQs 17(c) 2018 by Marcos Lopez de
Page 45 and 46: EXERCISES 19Dr. Lee Cohn, Dr. Micha
Page 49 and 50: INDEX 355(c) 2018 by Marcos Lopez d
Page 61: Praise forADVANCES in FINANCIAL MAC

SSRN-id3104847

Create successful ePaper yourself

Delete template?

Save as template?